api.last-price.ai
https://api.last-price.ai/api/mcp/rpc?profile=directory&via=openai-plugin
active Production-safe Score 93/100 http no auth
At a glance
| Endpoint | https://api.last-price.ai/api/mcp/rpc?profile=directory&via=openai-plugin |
|---|---|
| Source | https://github.com/last-price-ai/skills/blob/main/mcp.json |
| Transport | http |
| Tools | 308 |
| Auth | none |
| First seen | 2026-11-02 14:58:10 |
| Last checked | 2026-11-17 21:58:21 |
Observed reliability
| 7-day uptime | 100% |
|---|---|
| 24-hour uptime | 100% |
| Handshake success | 100% |
| Latency p50 / p95 | 543 / 907 ms |
| Last successful check | 2026-11-17 21:58:21 |
Measured by Arclan's validator via repeated live MCP handshakes. How this is computed →
Tools (308)
pricing_conversation— Ask Last Price what pricing inputs are missing. Start with request: {}, return its questions to the caller, then send continuation.request acompute_price— Compute a price. Returns the price, the function that produced it, the metered cost and a `trace` of each stage. Send currency plus unit_cosverify_price— Ask whether a price makes sense for its context, rather than what the price should be. Returns a verdict (within_band, too_high, too_low, orregister_agent— Sign up with no key and no human: creates a sandbox workspace of your own and returns its API key (shown once), its daily limits and a claimuse_api_key— Use a Last Price API key you already hold (a saved file, LASTPRICE_API_KEY) in this MCP connection, when it was opened without one: pass apimint_child_key— Create a short-lived child key for a sub-agent or job, returned only once. Ask for 1 to 20 concrete scopes within your current grants; no wirevoke_child_key— Revoke a direct child of the key this connection uses, by its key id. The child and its descendants stop working. You cannot revoke anotherrun_pricing_chain— Run a pricing chain: stages one after another, each stage's price feeding the next as current_price and its metadata carrying forward, yourlist_functions— List the pricing functions registered for the authenticated tenant.browse_catalog— Browse the public marketplace catalog of pricing functions.verify_signed_quote— Check a signed price quote you received from a seller: was it issued by Last Price, is it unaltered, and is it still valid. Pass its jws strget_offer_terms— Read a seller's negotiable terms by their public handle: which products are open to offers, at what list price and currency, how many offersmake_offer— Name a price for a seller's product and get the answer: accept, counter or reject, with the price and the reason. The first offer opens a neget_negotiation— Read your own negotiation with a seller: its status, the rounds used and left, the standing counter, the agreed price and the history. Needsaccept_counter— Take the seller's standing counter: the negotiation closes as accepted at that price. This commits the buyer to the price, so call it only wlist_pricing_pipelines— See the multi-step pricing pipelines Last Price publishes and the stages they are built from: for each reference pipeline its name, what itlist_hosted_demos— List the demo apps Last Price runs on its own site under /demos (a SaaS plan page, an agent bank, an AI API, a storefront, a freelancer markprice_from_capture— Price an item from what you have about it, the way a person would look at it: send photos, screenshots, PDFs, product page links, a descriptlist_price_checks— The workspace's price history: every price from a photo, link, PDF or text, newest first, whichever way it was asked (surface app for the phget_price_check— One check from the workspace's price history with its whole answer, as price_from_capture first gave it: the item and its condition, the prireport_problem— Tell Last Price a price or an answer is wrong, or send an idea, on the person's behalf: when they say a price from price_from_capture is offdescribe_item_so_far— While the person is still typing or talking about an item, see what it sounds like so far: send the words so far in text (they may stop mid-get_changelog— What changed in Last Price: the newest release notes, optionally for one feature, written for agents as well as people. Pass feature (a tagreport_outcome— Report whether a price from compute_price was used: once the buyer has accepted or turned down a price, report it. Pass receipt.id from thatget_price_guardrails— Read the price guardrails saved for your workspace, and the defaults set_price_guardrails fills in for an omitted field. saved is false whenset_price_guardrails— Save the price guardrails for your workspace. max_change_pct, min_margin_pct and never_below_cost hold every final price your workspace is gget_prompt_class_benchmarks— Per-prompt-class quality benchmarks across pricing functions and providers: for each kind of request (no text, code, extraction or JSON, mulget_sales_playbook— Read your workspace's sales playbook (Settings > Sales playbook) and what pricing and negotiation read from it: for each field, what was reaupdate_sales_playbook— Change some fields of your workspace's sales playbook and get back what pricing and negotiation now read from it. A field you send replacesget_channel_rates— Read the channel rates saved for your workspace: what each sales channel, such as a shopping agent or an agent marketplace, keeps of a sale,set_channel_rates— Save the channel rates for your workspace. A price request that names a saved channel (the channel argument of compute_price) is raised so yget_pricing_analysis— Read what your workspace's own records say about its pricing, as findings, most urgent first: learned estimates ready while learned pricingget_analyst_brief— Read the latest weekly brief the pricing analyst wrote for your workspace: short claims, each citing the findings it rests on (every numberask_pricing_analyst— Ask the pricing analyst one question about your workspace's pricing, in plain words, and get short claims that each cite the findings or estget_analyst_settings— Read whether the pricing analyst writes your workspace a brief, how often, whether a written brief may be paid from your Last Price balanceset_analyst_settings— Change the pricing analyst settings for your workspace: turn the brief on or off, choose weekly or nightly, choose whether a written brief mget_price_exploration— Read whether your workspace explores prices: learned and optimizer prices for the same product then vary slightly from call to call, by up tset_price_exploration— Turn price exploration on or off for your workspace, set its rate (at most 10% either way), or set how far past the prices already tried theget_frontier_model— Read the frontier price model: one network trained on millions of simulated markets across every vertical, pricing topology, unit, buyer andget_price_foundation_report— Read what the shadow ML pricing mix learned last night from your workspace's outcomes. On every learned quote it logs the price it would havsuggest_frontier_price— Ask the frontier price model what it would charge for one of your products: it reads your own quotes for that product from the last 90 daysget_buyer_aware_pricing— Read buyer-aware pricing for your workspace: models learned nightly from your own quotes (a contextual bandit and a boosted demand model, plset_buyer_aware_pricing— Set buyer-aware pricing for your workspace: the mode (off, shadow or live), which model prices in live mode, and which request fields the moget_feature_contract— Read your workspace's feature contract: which fields of a pricing request, nested metadata included, your buyer-aware pricing models learn fset_feature_contract— Save your workspace's feature contract, replacing the one saved: the request fields, by API path, your buyer-aware pricing models learn fromdelete_feature_contract— Remove your workspace's feature contract: quotes are logged without contract labels again. Needs a Last Price credential with routing writeinfer_feature_contract— Propose a feature contract from sample pricing requests: nested objects are flattened by path, lists reduced to counts, sums, maxima and hasget_offer_settings— Read how your workspace takes offers and counter-offers from buyers: whether offers are on, your public handle, the rounds, concession and eset_offer_settings— Turn offers and counter-offers on or off for your workspace and set the schedule every negotiation runs on. Omitted fields keep their savedset_offer_product— Open one product to offers, or change it: this saves the product whole (create or replace). Buyers see the list price and never the floor prremove_offer_product— Take a product off offers. Negotiations already open on it stop, because a product that is not listed answers buyers with not found. Needs alist_negotiations— Your workspace's negotiations, newest first: what a buyer sees plus the floor and asking price each one opened with. Use it to honour an accpreview_offer— Try an offer against one of your own products and see the answer a buyer would get (accept, counter or reject), with the limits behind it. Nget_function— Read one pricing function your workspace can use: its name, description, status, tags, cost per unit, cost multiplier, and input and outputtest_function— Test-run one pricing function with sample data and see the price, explanation, confidence and alternatives it returns, without publishing itlist_storefronts— List your workspace's storefronts: each one's subdomain, name, whether it is published, what it lists (functions, products open to offers, tget_storefront— Read one storefront with its domains, addresses and the offers it lists right now. Only where the workspace has storefronts switched on (elscreate_storefront— Create a storefront: an agent card, llms.txt and MCP endpoint listing a selection of your public offers at <subdomain>.store.last-price.ai (update_storefront— Change a storefront. Only the fields you send change; an empty string clears optional text. Set published to put it live or take it down. Ondelete_storefront— Delete a storefront and its domains. Its addresses stop answering at once. Only where the workspace has storefronts switched on (else an err
Validation history
| Checked | Reachable | Handshake | Latency | Score | State |
|---|---|---|---|---|---|
| 2026-11-17 21:58:21 | ✓ | ✓ | 908 ms | 90 | active |
| 2026-11-17 21:43:19 | ✓ | ✓ | 776 ms | 90 | active |
| 2026-11-17 21:28:17 | ✓ | ✓ | 539 ms | 90 | active |
| 2026-11-17 21:13:11 | ✓ | ✓ | 660 ms | 90 | active |
| 2026-11-17 20:57:34 | ✓ | ✓ | 607 ms | 90 | active |
| 2026-11-17 20:42:33 | ✓ | ✓ | 313 ms | 90 | active |
| 2026-11-17 20:27:32 | ✓ | ✓ | 299 ms | 94 | active |
| 2026-11-17 20:12:31 | ✓ | ✓ | 660 ms | 90 | active |
| 2026-11-17 19:55:14 | ✓ | ✓ | 797 ms | 90 | active |
| 2026-11-17 19:40:08 | ✓ | ✓ | 504 ms | 90 | active |
| 2026-11-17 19:24:48 | ✓ | ✓ | 342 ms | 90 | active |
| 2026-11-17 19:09:42 | ✓ | ✓ | 943 ms | 90 | active |
| 2026-11-17 18:54:34 | ✓ | ✓ | 654 ms | 90 | active |
| 2026-11-17 18:39:27 | ✓ | ✓ | 435 ms | 90 | active |
| 2026-11-17 18:24:20 | ✓ | ✓ | 498 ms | 90 | active |
Query this programmatically
Agents can pull this record live over the MCP protocol or REST:
get_server("https://api.last-price.ai/api/mcp/rpc?profile=directory&via=openai-plugin") # via the Arclan MCP endpoint https://arclan.ai/mcp
GET https://arclan.ai/api/registry/api-last-price-ai-api-mcp-rpc-profile-directory-via-openai-plugin # JSON